IgM anti-protein S antibodies as a risk factor for venous thrombosis

Lupus anticoagulant (LA) and antiphospholipid antibodies (APA) are immunoglobulins directed at phospholipid-protein complexes that have been associated frequently with thrombophilia. Protein S (PS) is a protein with high affinity for phospholipids and can be a target for APA. It is a cofactor in the Protein C (PC) anticoagulant pathway and facilitates the inactivation of coagulation Factors Va and VIIIa. We performed a case-control study to assess the prevalence of anti-protein S antibodies (anti-PS) in a Spanish population and to determine if they are a risk factor for venous thrombosis (VT).

Patients were included when they had been referred to our hospital from November 1997 to April 2002. This case-control study has been described previously in detail.1 Briefly: from the initial case-control study that included 250 patients and 250 controls, we obtained plasma for anti-PS analysis from 244 patients (108 males, 136 females) and 246 controls (107 males, 139 females). The patients were included in the study if they had suffered their first thrombotic event when under 70 years of age. Patients’ clinical characteristics are shown in Table 1. The control subjects were recruited according to the following criteria: similar age (±10 years), same sex, no genetic relationship to the patients and no personal or family history of VT. All procedures were approved by the Institutional Review Board of the Hospital de la Santa Creu i Sant Pau in Barcelona. Written informed consent was obtained from all participants.

Blood samples were obtained at least six months after the most recent thombotic event. IgG and IgM anti-PS antibodies were determined by an ELISA kit from Hyphen BioMed (Neuville-sur-Oise), specific for the IgG or IgM isotype. Normal range ≤8.2 AU for IgM and ≤8.9 AU for IgG (99 percentile of the distribution in the control population). The variation coefficient according to the manufacturer ranged from 3 to 6% for intra-assay and 4 to 8% for inter-assay. Tests for LA used Russell viper venom, (Life Diagnostics Frenchs Forest NSW Australia). IgG and IgM APA, including anticardiolipin, antiphosphatidylserine and anti-β2 Glycoprotein I were determined by ELISA methods. Antithrombin, PC, activated protein C resistance, total and free PS, Factor VIII clotting activity, Factor V Leiden (FVL), PT20210A of prothrombin gene, and F12C46T polymorphism were analyzed as previously described.1 Age was expressed as mean ± standard error (SE). The frequencies of variables are expressed in percentages and were compared with the χ test. A logistic regression method was used to estimate both the crude and adjusted Odds Ratio (OR) as a measure of risk. Adjustments were made for sex, age, LA and for other APA as covariables. Also, adjustments were made for those factors previously associated with VT in our population including: FVL, PT20210A, F12C46T (T/T), and levels of FVIII >90 percentile. The SPSS 14.0 software package was used for statistical analyses.

We found that 12 patients (4.9%) and 2 controls (0.8%) had IgM anti-PS antibodies. The associated thrombotic risk was 6.3 (95% CI: 1.4–28.5). After including LA and APA as covariables, the adjusted OR was 6.1 (95% CI: 1.3–28.0) (Table 2). The risk did not change when other thrombotic risk factors were included in the analysis (adjusted OR: 7.0; 95% CI: 1.5–33.5). After excluding patients with recurrent thrombosis, we studied 183 patients; 8 of them (4.4%) had IgM anti-PS. The associated risk was similar; the crude OR was 5.6 (95% CI:1.2–26.6). IgG anti-PS antibodies were present in 5 patients (2.0%) and in 2 controls (0.8%). The crude OR was 2.5 (95% CI: 0.5–13.3). Only 2 patients with IgM anti-PS presented LA or other APA. The remaining 10 patients and 2 controls with IgM anti PS did not present LA or other APA.

Table 1.Patients’ clinical characteristics.

Table 2.Thrombotic risk of anti-PS antibodies and of other thrombotic risk factors in our population.

Early reports2,3 have described anti-PS antibodies as complications of varicella associated with a transitory decrease of PS and thrombosis. Later, these antibodies were described in patients with systemic lupus erythematosus (SLE) that had suffered arterial or VT.4,5 Nojima et al.6 studied only the IgG anti-PS antibodies and reported that they constituted a strong risk factor for DVT also in non-SLE patients. Recently, Galli et al.7 reported that IgM anti-PS was associated with a high risk of prospective thrombosis (OR 6.58; 95% CI: 1.19–36.36) in the WAPS study (patients with persistent LA or APA). In our study, we found a 6-fold increased risk of thrombosis associated to IgM anti-PS and this risk was independent of LA, APA or other thrombotic risk factors.

The effect of these antibodies on hemostasis and their possible influence on thrombotic risk is still inconclusive. They may bind to the free form of PS and thereby inhibit its function8 or they may form immune complexes with PS causing an increased clearance of this protein.9 The possible link of these antibodies to thrombotic risk was suggested by Regnault et al.10 who reported one case in which, after varicella infection, there was a temporal relationship between PS deficiency, antibodies to PS and hypercoagulability in the Calibrated Thrombin Generation Test.

Our study has some limitations. First, we did not study the persistence of these antibodies in a second determination 12 weeks apart. However, since the blood of patients was obtained at least six months after their last thrombotic event, it is not likely that the antibodies that we detected were a transitory consequence of thrombosis. Second, since we are a reference center in our region, it is possible that the patients we studied represented a sample with higher thrombotic risk than a random group of patients. In this case, the thrombotic risk observed might be slightly lower in the general population. Finally, more studies of different populations would be desirable to confirm our results.

In conclusion, our study supports the hypothesis that IgM anti-PS antibodies could be an independent risk factor for thrombosis in a Spanish population.

Acknowledgments

we acknowledge the advice and helpful discussion of Professor W.H.Stone.

Footnotes

• Funding: this study was supported in part by the AGAUR (Agència de Gestió d’Ajuts Universitaris i de Recerca) Generalitat de Catalunya 2005 SGR 01068 and the Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III, RECAVA, (RD/0014/0016) and FIS 04/2165.